Background: Leukostasis is highly fatal. The entity is associated with hyperviscosity and microvascular ischemia. Risk factors include white blood cell count (WBC) >50K/uL, blasts CD56 expression and specific genomic abnormalities such as core binding factor (CBF), FLT3, RAS, among others. Our group recently confirmed that AML patients (pt) with leukostasis have dismal outcomes [30-day mortality; 50% v 14%, p=0.004, and overall survival (OS) of 158.4 d v 701 d, p=<0.0001] Fig 1A. Interestingly, adhesion and prothrombotic molecules' upregulation are hypothesized to underpin pathogenesis. However, characterization of Gene Expression (GE) deregulation is lacking, representing an unmet urgent need. In this study, we seek to validate previously known factors associated with leukostasis, and to uncover relevant biochemical predictors for the entity. Additionally, we aim to characterize deregulated gene pathways potentially inducing leukostasis.

Methods: After IRB approval, 101 AML pt were included for analysis. Phenotypic Leukostasis (PL) was defined as: (a) AML cases with >50K/uL WBC at presentation; (b) clinical evidence of respiratory distress, neurologic symptoms or detectable thrombotic or ischemic events [i.e., myocardial infarction, bowel or limb ischemia]. Demographic, clinical, laboratory and immunophenotypic AML characteristics were summarized in 32/101 (32%) and 69/101 (68%) of pt with and without leukostasis, respectively. Multivariate logistic model was developed to select variables from univariate analysis with strong predictive value.

Results: In pt with and without PL, mean age was 66 years (32-89) vs. 63.2 years (25-88) (p=NS), WBC was 83K/uL (1.6-351) vs. 28.4K/UL (0.6-270) (p=0.0024), absolute blast count was 47.2K/uL (0.07-275) vs. 17.3K/uL (0-195) (p= 0.01), platelet count was 60K/uL (6-278) vs. 101K/uL (3-531) (p=0.006), respectively. Evidence of tumor lysis syndrome (TLS) was observed in 14/32 (44%) and 10/67 (15%) of pt with and without PL, respectively (p=0.001). Monocytic progenitor (MP) immunophenotype was associated with PL, with a hazard ratio (HR) of 0.36 (95% CI 0.13-0.97, p=0.04) and when combining MP plus Granulocytic progenitor (GP) phenotype, HR=0.29 (CI 0.12-0.7, p=0.006). CD56 expression conferred a HR=0.22 (0.06-0.7, p=0.01), FLT3 mutation a HR= 0.15 (0.04-0.56, p=0.004) and there was a trend for NPM1 association (HR=0.08, CI 0.006-1.06, p=0.05). After adjusting for multiple variables, only CD56 expression (HR= -3.2, p= 0.009) and TLS (HR= -2.88, p=0.002) retained predictive value. Indeed, PL was predicted in 96% of cases presenting with CD56+ disease and TLS. As an exploratory aim, in pt with and without PL, we observed deregulated GE pathways in EDIL (endothelial cell behavior) 0.01 vs. 0.98, p=0.04, COL3A1 (collagen), 0.23 vs. 4.78 (p=0.05), PDL1, 2.72 vs. 8.5 (p=0.009) and IL1R, 2.72 vs 8.2 (p=0.05).

Conclusions: Our report demonstrates that in AML pt presenting with WBC>50K/uL, respiratory distress and/or suspected thrombotic/ischemic events, CD56 expression and TLS provides an accurate prediction of fatal PL. Abnormal transcriptomic networks leading to changes in endothelial behavior, AML niche collagen disruptions and deregulated immunity/inflammation signals potentially facilitate PL pathogenesis.

Disclosures

Renteria:Kymera Therapeutics: Consultancy; Novo Nordisk: Consultancy.

This content is only available as a PDF.
2024
Sign in via your Institution